Using a lattice structure coupon sample for build quality monitoring in metal additive manufacturing.

Simple item page
dc.contributor.advisorDu Plessis, Aen_ZA
dc.contributor.advisorVenter, MPen_ZA
dc.contributor.authorPark, Men_ZA
dc.contributor.otherStellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.en_ZA
dc.date.accessioned2023-11-22T18:17:48Zen_ZA
dc.date.accessioned2024-01-08T23:01:34Zen_ZA
dc.date.available2023-11-22T18:17:48Zen_ZA
dc.date.available2024-01-08T23:01:34Zen_ZA
dc.date.issued2023-11en_ZA
dc.descriptionThesis (MEng)--Stellenbosch University, 2023.en_ZA
dc.description.abstractENGLISH ABSTRACT: Additive manufacturing (AM), particularly Laser Powder Bed Fusion (L-PBF), has revolutionised the manufacturing industry by offering enhanced geometrical design freedom, part integration, and reduced production lead time. However, the presence of manufacturing defects during the L-PBF process poses challenges to the mechanical properties and build quality of the manufactured parts. Traditional quality assessment methods involve costly and time-consuming manufacturing, destructive or non-destructive analysis and mechanical testing. To address these challenges, this research proposes a novel approach for quality control using lattice coupon samples instead of solid mechanical test specimens. Lattice structures, comprising repetitive unit cells with struts and nodes, serve as effective indicators of build quality. The smaller size and volume of the lattice coupon samples result in significant cost and time savings compared to traditional test samples. The high sensitivity of lattices to parameter deviations enables their use in assessing build quality within an L-PBF system. These coupon samples provide reliable indicators of structural integrity and long-term performance of AM parts, simplifying the quality control process and optimising L-PBF manufacturing. Additionally, this research develops a Finite Element (FE) model for AM lattice structures under quasi-static compression, offering a powerful tool for the virtual testing of components, enabling the study of their mechanical behaviour without the need for costly physical prototypes. The FE model incorporates defect states of the AM lattice structures observed from computed tomography (CT) scanning. By taking input from the CT scans, the prediction of mechanical properties of lattices with a high accuracy rate of 83 % was achievable. This model represents a promising tool for developing manufacturing defect-incorporated lattice representative volume elements (RVEs) for use in the design of AM parts incorporating lattice regions. By replacing complex lattice structures with solid-infilled features in the form of RVEs in simulations, the computational expense can be significantly reduced. This approach allows for efficient exploration of the mechanical behaviour of latticed AM components while accounting for manufacturing defects, offering insights for design optimisation and material selection. Furthermore, this research aims to leverage the developed FE model and interpolation methods to predict the mechanical properties of lattice structures, reducing the reliance on physical printing and CT scanning. By utilising these computational tools, accurate estimations of the mechanical properties can be achieved, minimising the need for extensive experimental testing and CT scan. Overall, this research contributes to the advancement of quality control in AM by introducing lattice coupon samples as indicators of build quality and developing computational models for predicting their mechanical properties. These innovations lead to improved efficiency and optimisation of L-PBF manufacturing processes, benefiting industries that rely on AM technology.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: Laagvervaardiging (LV), veral Laser Poeier Bed Samesmelting (L-PBS), het die vervaardigingsbedryf gerevolusioneer deur verbeterdeontwerpvryheid, integrasie van onderdele en verminderde produksie-aflewertyd aan te bied. Die teenwoordigheid van vervaardigingsfoute tydens die L-PBS-proses stel egter uitdagings aan die meganiese eienskappe en bougehalte van die vervaardigde onderdele. Tradisionele kwaliteitsbeoordelingsmetodes behels koste en tydrowende vervaardiging, vernietigende of nie-vernietigende analise, en meganiese toetse. Om hierdie uitdagings aan te spreek, stel hierdie navorsing 'n nuwe benadering voor vir kwaliteitsbeheer deur gebruik te maak van rasterkoeponmonsters (oftewel traliestruktuur) in plaas van soliede meganiese toetsmonsters. Rasterstrukture, bestaande uit herhalende eenheidselle met stut en nodusse, dien as doeltreffende aanduiders van bougehalte. Die kleinere grootte en volume van rasterkoeponmonsters lei tot aansienlike koste- en tydbesparings in vergelyking met tradisionele toetsmonsters. Die hoë sensitiwiteit van rasterstrukture vir afwykings in parameters maak dit moontlik om bougehalte binne 'n L-PBS-sisteem te evalueer. Hierdie koeponmonsters bied betroubare aanduiders van die strukturele integriteit en langtermynprestasie van LV-onderdele, wat die kwaliteitsbeheerproses vereenvoudig en die L-PBS-vervaardigingsproses optimeer. Daarbenewens ontwikkel hierdie navorsing 'n Eindige Element (EE) model vir LV-rasterstrukture onder quasi-statiese druk, wat 'n kragtige hulpmiddel bied vir die virtuele toetsing van komponente en die studie van hul meganiese gedrag sonder die noodsaak van duur fisiese prototipes. Die EE-model inkorporeer defektoestande van die LVrasterstrukture wat waargeneem is vanuit rekonstrueerde tomografie (RT) skandering. Deur inligting van die RT-skandering te gebruik, was die voorspelling van meganiese eienskappe van die raster met 'n hoë akkuraatheid van 83 % haalbaar. Hierdie model verteenwoordig 'n belowende hulpmiddel vir die ontwikkeling van vervaardigingsfout-geïnkorporeerde verteenwoordigende volumeelemente (VVEs) van die raster vir die ontwerp van LV-onderdele wat rasterareas bevat. Deur komplekse rasterstrukture te vervang met vol-gedeelde kenmerke in die vorm van VVEs in simulasies, kan die berekeningskoste aansienlik verminder word. Hierdie benadering maak effektiewe ondersoek van die meganiese gedrag van gerasterde LV-komponente moontlik terwyl vervaardigingsdefekte in ag geneem word, en bied insigte vir die optimalisering van ontwerp en materiaalkeuse. Verder beoog hierdie navorsing om die ontwikkelde EE-model en interpolasiemetodes te gebruik om die meganiese eienskappe van rasterstrukture te voorspel, wat die afhanklikheid van fisiese druk en RT-skandering verminder. Deur hierdie berekeningshulpmiddels te gebruik, kan akkurate skattings van die meganiese eienskappe bereik word, terwyl die behoefte aan omvangryke eksperimentele toetse verminder word. Hierdie benadering verminder aansienlik koste, tyd en hulpbronne, terwyl dit steeds waardevolle insigte bied in die meganiese gedrag van rasterstrukture. Alles in ag genome, dra hierdie navorsing by tot die bevordering van kwaliteitsbeheer in LV deur die gebruik van rasterkoeponmonsters as aanduiders van bougehalte en die ontwikkeling van berekeningsmodelle vir die voorspelling van meganiese eienskappe. Hierdie innovasies lei tot verbeterde doeltreffendheid en optimisering van L-PBSvervaardigingsprosesse, wat voordelig is vir nywerhede wat op LV-tegnologie staatmaak.af_ZA
dc.description.versionMastersen_ZA
dc.format.extent110 pages : illustrationsen_ZA
dc.identifier.urihttps://scholar.sun.ac.za/handle/10019.1/129117en_ZA
dc.language.isoen_ZAen_ZA
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subject.lcshAdditive manufacturingen_ZA
dc.subject.lcshGeometric patternsen_ZA
dc.subject.lcshLean manufacturingen_ZA
dc.titleUsing a lattice structure coupon sample for build quality monitoring in metal additive manufacturing.en_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
park_using_2023.pdf
Size:
5.23 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Masters Degrees (Mechanical and Mechatronic Engineering)